XJTLU Learning Mall

Introduction

The mastery of academic writing is widely recognized as one of the most crucial skills in the academic setting, yet it is often also the most challenging for students to acquire. At XJTLU, the emphasis on teaching academic writing is throughout the EAP classes. Despite this, EAP instructors frequently encounter difficulties in crafting writing resources. For example, teachers have to manage multiple tasks, such as lesson planning, marking and service duties, within a limited time frame, which may potentially diminish the time available for creating writing practice materials.

XIPU AI is an artificial intelligence system deployed by XJTLU. The major advantage of it is that it can generate text in response to a given prompt, such as a question or statement. This characteristic enables it a significant advantage in personalized content creation, presenting a valuable resource for assisting EAP teachers in overcoming their challenges. Specifically, it can be used to generate model paragraphs or essays that serve as examples of high-quality writing, as well as creating samples with different types of errors, which students can practice identifying and amending. This not only effectively saves time and effort of teachers, but also helps to improve writing skills of students. This article will explore the potential applications of XIPU AI in producing both exemplars and non-exemplars for EAP writing class.

Generating Exemplars

According to Sadler (2005, p.192), exemplars are “key examples chosen so as to be typical of designated levels of quality or competence”. In education, exemplars represent ideas or concept that a teacher is trying to teach their students. They are considered as crucial pedagogical tools that help learners form accurate concept (To, Panadero & Carless, 2022). Specifically, exemplars work as standard frameworks for students to interpret assessment criteria and understand teachers' expectations of performance. Students are also able to reflect on their own work after comparing with exemplars and then make improvements (To, Panadero & Carless, 2022).

For EAP writing coursework, students are often provided with a model essay which was crafted by their teachers. However, students may prefer to be exposed to more exemplars, which can augment their comprehension of the essay genre, as well as necessary components and favorable attributes of the essay. Teachers also seek additional samples to use as instructional reference when teaching specific essay features. XIPU AI is an effective tool that can help generate more exemplars. Figure 1 is an example of an introduction paragraph of a Problem-Solution-evaluation (PSE) essay generated by XIPU AI.

Figure 1 an exemplar introduction paragraph of a PSE essay generated by XIPU AI

In order to achieve optimal outcomes, numerous attempts have been made to optimize the prompt, with the above version being the most aligned with expectations. Consequently, several key factors should be considered when formulating prompts. Firstly, it is essential to provide clear context within the instructions. This includes specifying the essay genre and structure (PSE essay), the main idea of the paragraph (an introduction), language level (CEFR B2), and the desired answer length (approximately 100 words). XIPU AI has demonstrated enhanced performance when provided with such contextual information. Secondly, instructors should specify the critical information to be incorporated. For instance, the introduction paragraph should consist of a specified numbers of sentences addressing various aspects. All this provided descriptive information aligns with what has been described in students’ writing coursework task sheet, which significantly enhances the likelihood of receiving accurate responses. Finally, the provision of an example paragraph, as illustrated in Figure 2, can further assist in guiding XIPU AI. This example, which was excerpted from an existing model essay, enables XIPU AI to emulate the style of the expected answer and generate precisely what the instructor requires.

Figure 2 an exemplar solution paragraph of a PSE essay generated by XIPU AI

Generating non-exemplars

Non-exemplars, in contrast to exemplars, serve to highlight points of difference rather than similarity. While exemplars provide clear examples of a concept, non-exemplars, although sharing some features with the concept, exhibit one or more differing attributes, thereby delineating the boundaries of the concept. This enables students to figure out the relevant characteristics of the newly acquired concept, refining their understanding of its defining features. By exposing learners to a wide range of non-exemplars, educators allow them to critically evaluate the concept, as students can reflect on and articulate why the non-exemplars fail to represent the concept, thus deepening their ability to distinguish between exemplars and non-exemplars (Metropolitan State University of Denver, 2022).

In EAP classes, teachers often share non-exemplars of the writing task in class to facilitate students’ analysis and comparison, thereby enhance their comprehension of important essay components. This pedagogical approach serves as an additional scaffold for EAP learners. Figure 3 shows both an exemplar and a non-exemplar of an argumentative paragraph generated by XIPU AI. In the non-exemplar, less explanations and examples are involved in supporting details. By comparing these two versions, students can better appreciate the critical that reasons and examples play in idea development in a paragraph.

Figure 3 an exemplar and a non-exemplar of an argumentative paragraph generated by XIPU AI

Figure 4 illustrates another non-exemplar generated by XIPU AI. Much of the vocabulary in the previously generated solution paragraph is replaced with less advanced and academic ones. This illustration can serve as a valuable resource for classroom pair discussions, enabling students to discern and analyze the differences, or as an exercise to encourage them to enhance their vocabulary skills.

Figure 4 a non-exemplar of a solution paragraph of a PSE essay generated by XIPU AI

Reflection

Leveraging XIPU AI to generate both exemplars and non-exemplars for EAP writing confers benefits both educators and learners. A primary advantage for teachers lies in the reduction of their workload. By harnessing AI capabilities, teachers are able to automate certain tasks associated with teaching materials designing, thereby affording them additional time to focus on other aspects of their job, such as offering constructive feedback to students and creating novel instructional resources (Mondal et al, 2023).

One of the advantages for students is increased clarity regarding coursework requirements and standards, facilitated by the ability to compare and practice a wide range of positive and negative examples generated by AI. To, Panadero, and Carless (2015, p.1169) found that “students generally perceived exemplars to be effective in clarifying assessment task standards, guiding task preparation and increasing their understanding of quality.” A second benefit is enhanced academic performance among students. With a clear understanding of coursework requirements, students cultivate heightened confidence and self-efficacy in their task engagement (To and Carless, 2016). The writer has also observed an improvement in writing performance after students participated in exemplar and non-exemplar grading and discussion.

Despite benefits associated with the use of AI in exemplar and non-exemplar generation, it is important to acknowledge the presence of challenges. Firstly, due to the inherent limitations of AI, it is unable to reason and interpret context beyond the current sentences or paragraphs, so its outputs may not consistently align with the specific requirements of various EAP modules. Consequently, teachers may sometimes encounter instances where the responses fail to precisely meet their expectations. Therefore, it is recommended to establish enhanced dialogues between users and AI, and to continually refine prompts based on the situation to achieve the optimal outcome. Furthermore, teachers should be capable of editing AI’s answers to provide tailored exemplars and non-exemplars for students to practice.

Another potential concern associated with the utilization of AI for text generation is student plagiarism. Given AI’s capability in providing answers based on instructions, students can also request sample paragraphs, thereby exacerbating the risk of plagiarism. This not only compromises academic integrity but also runs counter to the purpose of assessment, which is to fairly evaluate student learning. Some strategies can be considered to combat this issue. One possible way is the incorporation of specified sources within paragraphs, as AI is unable to integrate specific sources to support ideas. Moreover, instructor training and student education are critical. For example, educators should be provided with the requisite training to identify instances of AI use in student assignments, which can be facilitated through the use of AI detection tools like Turnitin. Equally important is increasing students’ awareness of academic integrity policies and ensuring their understanding of the consequences of academic misconduct (Lo, 2023).

Conclusion

The integration of XIPU AI has proven to be a valuable tool for EAP teachers, offering robust support in the generation of both exemplars and non-exemplars in writing practices. Leveraging XIPU AI accelerates the process of material development and save teachers’ time and efforts for other activities. The use of exemplars and non-exemplars plays pivotal part in students’ acquisition of academic writing skills, allowing them to better understand writing coursework standards and further improve academic performance in writing. However, AI’s inherent limitations and potential AI-lead cheating are two major challenges when harnessing XIPU AI for content generation. To address these issues, it is recommended specific strategies be devised. A comprehensive understanding of these capabilities and limitations will enable educators to optimize the potential of XIPU AI while effectively mitigating associated risks and challenges.

References

Lo, C.K. (2023) ‘What Is the Impact of ChatGPT on Education? A Rapid Review of the Literature’, Education Science, 13(410), Available at: https://doi.org/10.3390/educsci13040410 (Accessed: Aug 22 2014)

Metropolitan State University of Denver (2022) Teaching by example and nonexample
Available at： https://www.msudenver.edu/early-bird/teaching-by-example-and-nonexample/ (Accessed: July 29 2014)

Mondal, H., Marndi, G., Behera, J.K. & Mondal, S. (2023) ‘ChatGPT for Teachers: Practical Examples for Utilizing Artificial Intelligence for Educational Purposes’, Indian Journal of Vascular and Endovascular Surgery, doi: 10.4103/ijves.ijves_37_23
Sadler, R. (2005) ‘Interpretations of Criteria-Based Assessment and Grading in Higher Education’, Assessment and Evaluation in Higher Education, 30 (2), pp.175-194. doi: 10.1080/0260293042000264262.

To, J. & Carless, D. (2016) ‘Making Productive Use of Exemplars: Peer Discussion and Teacher Guidance for Positive Transfer of Strategies.’ Journal of Further and Higher Education, 40 (6), pp. 746-764. doi:10.1080/0309877X.2015.1014317.

To, J., Panadero, E. & Carless, D. (2022) ‘A systematic review of the educational uses and effects of exemplars’, Assessment & Evaluation in Higher Education, 47(8), pp.1167–1182. Available at: https://doi.org/10.1080/02602938.2021.2011134. (Accessed: July 29 2014)